R Notebook

Data Visualization Post

For portfolio at willdesi.com

library(tidyverse)
library(leaflet)

# below data downloaded from https://www.motherjones.com/politics/2012/12/mass-shootings-mother-jones-full-data/
df = read_csv("Mother Jones - Mass Shootings Database, 1982 - 2023 - Sheet1.csv")

df = df %>% mutate(gender = case_when(
    gender %in% "F" ~ "Female",
    gender %in% "M" ~ "Male",
    TRUE ~ gender))

df = df %>% mutate(race = case_when(race %in% "-" ~ "Unknown",
                                    race %in% "black" ~ "Black",
                                    race %in% "unclear" ~ "Unknown",
                                    TRUE ~ race))
# separate the location column into two columns; city and state. Keep the location column
df = df %>% tidyr::separate(col = location, into = c("city", "state"), sep = ", ", remove = FALSE)

# get the lat and long for the city and state
df = df %>% geocode(city = city, state = state)

# some lat and long values are NA where the original latitude and longitue are not null. So replace the corresponding NA values in lat and long with those in the other columns.
df = df %>% mutate(Char1 = coalesce(lat, latitude)) %>%
    mutate(Char2 = coalesce(long, longitude))

# remove unwanted columns and rename columns ilke location...2 to location
df = df %>% select(-latitude, -longitude, -lat, -long) %>%
    rename(location = location...2, latitude = Char1, longitude = Char2)

# save this cleaned data
write_csv(df, "mother_jones_mass_shootings_updated.csv")

DT::datatable(df, rownames = F, filter = "top", options = list(pageLength = 2, scrollX = TRUE
                                                               )) 

Maps

m = leaflet(df)
m %>%
    setView(lat = 37.09024, lng = -95.712891, zoom = 4) %>%
    addProviderTiles("Esri.WorldStreetMap") %>%
    #addProviderTiles(providers$CartoDB.Positron) %>%
    addCircles(radius = df$fatalities*1000, 
               fillOpacity = 0.25,
               popup = paste0("<strong>Case: </strong>", df$case, "<br>",
                              "<strong>Date: </strong>", df$date, "<br>",
                              "<strong>Place: </strong>", df$location, "<br>",
                              "<strong>Summary: </strong>", df$summary, "<br>",
                              "<strong>Fatalities: </strong>", df$fatalities, "<br>"))

Boxplots

Data for the plots

df1 = chickwts %>% mutate(sample = str_c("chick_", 1:n()))
my_colors = paletteer::paletteer_d("ggthemes::excel_Headlines")
head(df1)

df1 = chickwts %>% mutate(sample = str_c("chick_", 1:n()))
my_colors = paletteer::paletteer_d("ggthemes::excel_Headlines")

p2 <- ggplot(data = df1, aes(x = feed, y = weight, color = feed)) +
    geom_boxplot(outlier.shape = NA) + 
    geom_jitter()+
    ggthemes::scale_color_calc() +
    ggthemes::theme_fivethirtyeight() +
    labs(title = "Chicken Weights by Feed Type",
         x = "", y = "Weight in Grams")
p2

p <- ggplot(data = df1, aes(x = feed, y = weight, color = feed,
                            text = paste0("Sample: ", sample,
                                          "<br>","Weight: ",
                                          weight, "g"))) +
    geom_boxplot(outlier.shape = NA) + 
    geom_jitter() +
    scale_color_manual(values = my_colors) +
    theme_classic() +
    labs(title = "Chicken Weights by Feed Type",
         x = "", y = "Weight in Grams")

ggplotly(p, tooltip = "text") %>% 
    layout(legend = list(orientation = "h"))

fig <- plot_ly(df1, y = ~weight, color = ~feed, type = "box",
               boxpoints = "all", jitter = 0.3, pointpos = -1.8, 
               text = paste0("Sample: ", df1$sample, "<br>", "Weight: ", df1$weight, "g"),
               hoverinfo = 'text')

fig  %>% 
    layout(legend = list(orientation = "h"),
           title = list(text = "Chicken Weights by Feed Type"),
           yaxis = list(title = "Weight(grams)"))

Bar plot

# we first have to compute summary statistics of the data
df_summary = chickwts %>%
    group_by(feed) %>%
    summarise(mean = mean(weight),
              sd = sd(weight, na.rm = TRUE))

ggplot(df_summary, aes(x = feed, y = mean)) +
    geom_col(color = "black", fill = my_colors) +
    geom_errorbar(aes(ymin = mean, ymax=mean + sd), width = 0.2) +
    labs(title = "Chicken Weights by Feed Type",
         x = "", y = "Weight in Grams")

Violin Plots

# starting with a ggplot object
g <- ggplot(data = df1, aes(x = feed, y = weight, color = feed)) +
    scale_color_manual(values = my_colors) +
    scale_fill_manual(values = my_colors)

g2 = g +
    geom_violin(
        aes(fill = feed, fill = after_scale(colorspace::lighten(fill, .5))), size = 1) +
    geom_boxplot(
        fill = "white",  size = 1, width = .2, outlier.shape = NA, coef = 0) +
    geom_point(
        position = position_jitter(width = .03, seed = 0), size = 2, alpha = .5) +
    geom_point(
        position = position_jitter(width = .03, seed = 0), size = 2, stroke = .7, shape = 1, color = "black") +
    theme(legend.position = "none") +
    labs(title = "Chicken Weights by Feed Type", x = "", y = "Weight in Grams")
    
g2

fig <- df1 %>% plot_ly(x = ~feed, y = ~weight, split = ~feed, type = 'violin', box = list(visible = T), points = "all",
                       text = ~paste("Sample : ", sample , "<br>Weight: ", weight, "g")) %>%
    layout(legend = list(orientation = "h"),
           title = list(text = "Chicken Weights by Feed Type"),
           yaxis = list(title = "Weight in Grams"))
fig

Raincloud Plots

library(PupillometryR)
my_colors = paletteer::paletteer_d("ggthemes::excel_Headlines")

gg_rain = ggplot(data = df1, aes(x = feed, y = weight, fill = feed, color = feed)) +
    PupillometryR::geom_flat_violin(position = position_nudge(x = .2), alpha = .4, trim=FALSE) +
    gghalves::geom_half_point(side = "l", range_scale = .3, alpha = .5, size = 3) +
    geom_boxplot( width = .2, size = 1, outlier.shape = NA, fill = "white",) +
    coord_flip() +
    scale_color_manual(values = my_colors) +
    scale_fill_manual(values = my_colors) +
    ggthemes::theme_calc() +
    theme(legend.position = "none") +
    labs(title = "Chicken Weights by Feed Type", x = "", y = "Weight in Grams")
gg_rain

g3 = df1 %>% ggplot(aes(x = feed, y = weight, fill = feed)) +
    # the half violin
    ggdist::stat_halfeye(
    adjust = .5, ## bandwidth
     .width = c(0.66, 0.95), 
    position = position_nudge(x = .2), # move geom to the right
    color = NA, # removes the slab interval
    scale = 0.5
  ) +
    # the boxplot
    geom_boxplot(width = .2, size = 1, outlier.color = NA, alpha = 0.5) +
    # the dots
    ggdist::stat_dots(position = "dodge", scale = 0.5, side = "left", dotsize = 1, justification = 1.2) +
    ggthemes::theme_calc() +
    theme(legend.position = "none") +
    scale_fill_manual(values = my_colors) +
    labs(title = "Chicken Weights by Feed Type", x = "", y = "Weight in Grams") +
    coord_flip()
g3